Testing GPT and Claude on century-old geometry problems: proofs without seeing figures
snowboat84 · x · 2026-10-15
The author tested GPT and Claude on problems from a century-old Chinese geometry textbook (Yan Jici's method). Findings: some problems were solved correctly without drawing; others failed at key geometric relations with confident wrong proofs; models strongly prefer analytic/vector tools over synthetic geometry. This suggests LLMs still lack genuine spatial construction ability, raising the question of whether text-only Transformers "see" geometric configurations or rely purely on symbolic pattern matching — unlike AlphaGeometry's hybrid LLM-plus-symbolic-prover approach.
More from Models
- ChatGPT web can actually listen to audio, but the same model can't via API or Codex — Possible_Statement84 · 2026-10-23
- MiniMax M3.1 Flash free for 3 days on 0G, with 1M context and 512K output — MiniMax_AI · 2026-10-23
- TLA+ expert corrects Grok: TLC only exhausts finite state spaces, not all behaviors — tianyin_xu · 2026-10-23
- Open LLM cost calculator shows Arabic tokenization and 6x price gaps across models — Moham-saed · 2026-10-23
- OpenAI blocks manual reset with 13% usage left: 'no reset needed right now' — AccomplishedSugar490 · 2026-10-23
- Meta's 816B Muse Spark Open scores 39, losing to 180B Qwen3.8-Flash-Next at 40 — BTA_Labs · 2026-10-23